All of lore.kernel.org
 help / color / mirror / Atom feed
From: David Howells <dhowells@redhat.com>
To: Steve French <smfrench@gmail.com>
Cc: dhowells@redhat.com, Shyam Prasad N <nspmangalore@gmail.com>,
	Rohith Surabattula <rohiths.msft@gmail.com>,
	Jeff Layton <jlayton@kernel.org>,
	Al Viro <viro@zeniv.linux.org.uk>,
	linux-cifs@vger.kernel.org, linux-fsdevel@vger.kernel.org,
	linux-kernel@vger.kernel.org
Subject: [PATCH 1/6] iov_iter: Add a function to extract an iter's buffers to a bvec iter
Date: Mon, 09 May 2022 17:36:14 +0100	[thread overview]
Message-ID: <165211417421.3154751.5516678732460554467.stgit@warthog.procyon.org.uk> (raw)
In-Reply-To: <165211416682.3154751.17287804906832979514.stgit@warthog.procyon.org.uk>

Copy cifs's setup_aio_ctx_iter() and to lib/iov_iter.c and generalise it as
extract_iter_to_iter().  This allocates and sets up an array of bio_vecs
for all the page fragments in an I/O iterator and sets a second supplied
iterator to bvec-type pointing to the array.

This is can be used when setting up for a direct I/O or an asynchronous I/O
to set up a record of the page fragments that are going to contribute to
the buffer, paging them all in to prevent DIO->mmap loops and allowing the
original iterator to be deallocated (it may be on the stack of the caller).

Note that extract_iter_to_iter() doesn't actually need to make a separate
allocation for the page array.  It can place the page array at the end of
the bvec array storage, provided it traverses both arrays from the 0th
element forwards.

Signed-off-by: David Howells <dhowells@redhat.com>
---

 include/linux/uio.h |    4 ++
 lib/iov_iter.c      |   93 +++++++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 97 insertions(+)

diff --git a/include/linux/uio.h b/include/linux/uio.h
index 739285fe5a2f..5a3c6f296b96 100644
--- a/include/linux/uio.h
+++ b/include/linux/uio.h
@@ -236,6 +236,10 @@ ssize_t iov_iter_get_pages(struct iov_iter *i, struct page **pages,
 			size_t maxsize, unsigned maxpages, size_t *start);
 ssize_t iov_iter_get_pages_alloc(struct iov_iter *i, struct page ***pages,
 			size_t maxsize, size_t *start);
+ssize_t extract_iter_to_iter(struct iov_iter *orig,
+			     size_t orig_len,
+			     struct iov_iter *new,
+			     struct bio_vec **_bv);
 int iov_iter_npages(const struct iov_iter *i, int maxpages);
 void iov_iter_restore(struct iov_iter *i, struct iov_iter_state *state);
 
diff --git a/lib/iov_iter.c b/lib/iov_iter.c
index 6dd5330f7a99..8db34ddd23be 100644
--- a/lib/iov_iter.c
+++ b/lib/iov_iter.c
@@ -1696,6 +1696,99 @@ ssize_t iov_iter_get_pages_alloc(struct iov_iter *i,
 }
 EXPORT_SYMBOL(iov_iter_get_pages_alloc);
 
+/**
+ * extract_iter_to_iter - Extract the pages from one iterator into another
+ * @orig: The original iterator
+ * @orig_len: The amount of iterator to copy
+ * @new: The iterator to be set up
+ * @_bv: Where to store the allocated bvec table pointer, if allocated
+ *
+ * Extract the page fragments from the given amount of the source iterator and
+ * build up a second iterator that refers to all of those bits.  This allows
+ * the original iterator to disposed of.
+ *
+ * If a bvec array is created, the number of pages in the array is returned and
+ * a pointer to the array is saved into *@_bv;
+ */
+ssize_t extract_iter_to_iter(struct iov_iter *orig,
+			     size_t orig_len,
+			     struct iov_iter *new,
+			     struct bio_vec **_bv)
+{
+	struct bio_vec *bv = NULL;
+	struct page **pages;
+	unsigned int cur_npages;
+	unsigned int max_pages = iov_iter_npages(orig, INT_MAX);
+	unsigned int npages = 0;
+	unsigned int i;
+	size_t count = orig_len;
+	ssize_t ret;
+	size_t bv_size, pg_size;
+	size_t start;
+	size_t len;
+
+	*_bv = NULL;
+
+	if (iov_iter_is_kvec(orig) || iov_iter_is_discard(orig)) {
+		*new = *orig;
+		iov_iter_advance(orig, count);
+		return 0;
+	}
+
+	bv_size = array_size(max_pages, sizeof(*bv));
+	bv = kvmalloc(bv_size, GFP_KERNEL);
+	if (!bv)
+		return -ENOMEM;
+
+	/* Put the page list at the end of the bvec list storage.  bvec
+	 * elements are larger than page pointers, so as long as we work
+	 * 0->last, we should be fine.
+	 */
+	pg_size = array_size(max_pages, sizeof(*pages));
+	pages = (void *)bv + bv_size - pg_size;
+
+	while (count && npages < max_pages) {
+		ret = iov_iter_get_pages(orig, pages, count, max_pages - npages,
+					 &start);
+		if (ret < 0) {
+			pr_err("Couldn't get user pages (rc=%zd)\n", ret);
+			break;
+		}
+
+		if (ret > count) {
+			pr_err("get_pages rc=%zd more than %zu\n", ret, count);
+			break;
+		}
+
+		iov_iter_advance(orig, ret);
+		count -= ret;
+		ret += start;
+		cur_npages = DIV_ROUND_UP(ret, PAGE_SIZE);
+
+		if (npages + cur_npages > max_pages) {
+			pr_err("Out of bvec array capacity (%u vs %u)\n",
+			       npages + cur_npages, max_pages);
+			break;
+		}
+
+		for (i = 0; i < cur_npages; i++) {
+			len = ret > PAGE_SIZE ? PAGE_SIZE : ret;
+			bv[npages + i].bv_page	 = *pages++;
+			bv[npages + i].bv_offset = start;
+			bv[npages + i].bv_len	 = len - start;
+			ret -= len;
+			start = 0;
+		}
+
+		npages += cur_npages;
+	}
+
+	*_bv = bv;
+	iov_iter_bvec(new, iov_iter_rw(orig), bv, npages, orig_len - count);
+	return npages;
+}
+EXPORT_SYMBOL(extract_iter_to_iter);
+
 size_t csum_and_copy_from_iter(void *addr, size_t bytes, __wsum *csum,
 			       struct iov_iter *i)
 {



  reply	other threads:[~2022-05-09 16:36 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-05-09 16:36 [PATCH 0/6] cifs: Use iov_iters down to the network transport David Howells
2022-05-09 16:36 ` David Howells [this message]
2022-05-09 16:36 ` [PATCH 2/6] iov_iter: Add a general purpose iteration function David Howells
2022-05-09 16:36 ` [PATCH 3/6] cifs: Add some helper functions David Howells
2022-05-09 16:36 ` [PATCH 4/6] cifs: Add a function to read into an iter from a socket David Howells
2022-05-09 16:36 ` [PATCH 5/6] cifs: Change the I/O paths to use an iterator rather than a page list David Howells
2022-05-09 16:36 ` [PATCH 6/6] cifs: Remove unused code David Howells
2022-05-13 15:59 ` [PATCH 7/6] cifs: Make the cifs RDMA code use iterators David Howells

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=165211417421.3154751.5516678732460554467.stgit@warthog.procyon.org.uk \
    --to=dhowells@redhat.com \
    --cc=jlayton@kernel.org \
    --cc=linux-cifs@vger.kernel.org \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=nspmangalore@gmail.com \
    --cc=rohiths.msft@gmail.com \
    --cc=smfrench@gmail.com \
    --cc=viro@zeniv.linux.org.uk \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.